ActiveCampaign

Growth Today built an account research and scoring system in Clay that reads 15 custom data points per account and delivers the score inside Apollo, where the SDR team already works.

Impact TL;DR

• 50,000 accounts scored, covering the full target market rather than just a portion of target account
• 15 data points per account resolved automatically, replacing roughly twenty minutes of manual research each
• A single numeric score per account, with a working threshold of 35 to 40 for priority accounts
• Custom industry classification built from two inputs and mapped to ActiveCampaign’s own internal taxonomy
• Competitor scores, weighted by how well ActiveCampaign historically displaces each incumbent
• Scores delivered inside Apollo, so SDRs get the data where they already work

‍

A large TAM with no way to rank it

ActiveCampaign had the accounts and the headcount. What the SDR team lacked was any signal about which accounts deserved effort. An SDR opening an account record had no indication of its size, its stack, its growth trajectory, or whether ActiveCampaign had ever won against the tool it was already running.

So they researched by hand. Around a dozen data points per account, pulled from separate sources, at 10-20 minutes an account. Scoped as a manual project across the full 50,000 accounts, that path required four SDRs working full time for about a year, and it would have produced a snapshot that started decaying the day it finished.

The scoring problem also had a taxonomy problem inside it. ActiveCampaign classifies verticals in their CRM using their own framework, which does not map cleanly onto LinkedIn’s industry tags. Any classification built on the LinkedIn tag alone would have produced categories the sales team could not act on.

Key challenges were:

• No clear and actionable account prioritization available to SDRs at the point of work
• Roughly ten-twenty minutes of manual research per account per SDR
• A dozen separate sources to check, with no single record of the result to review, update and govern
• LinkedIn industry tags did not match ActiveCampaign’s internal vertical categorization taxonomy
• Any manual research data would go out of date, since traffic, headcount, and tech stack change continuously

‍

One reusable enrichment process that scores every account in less than 5 minutes in the full TAM

Growth Today built one reusable enrichment process in Clay that resolves every scoring input per account, sums them into a score, and pushes the result into Apollo.

Instead of asking SDRs to research accounts, we wanted to build a system where:

• Every account in the TAM has a highly accurate, clear score, not just the accounts someone got to
• Each score is composed of inputs the sales team already believes predict a win
• Competitor presence is weighted by displacement history rather than treated as one signal
• Industry classification matches with ActiveCampaign’s own Sales Team categories
• The output appears in Apollo, so the workflow does not change for the rep
• The whole research automation runs against the assigned TAM list in hours, instead of months of manual research, saving minimum 5 hours per SDR per week

None of this was going to work without agreement on what actually predicts a win. So we built the score from inputs ActiveCampaign’s team could defend, including:

• Monthly website traffic, which is a proxy for how many email subscribers a brand could have
• Company revenue, resolved through a revenue waterfall enrichment process
• Tech stack, split into competitor presence and integration presence, scored separately
• Headcount, headcount trend over 12 months, sales headcount, and open roles
• Aggregated review volume and aggregated social following

‍

Classifying industry against ActiveCampaign’s own framework

What we did

We took two inputs per account. The first was the industry tag on the company’s LinkedIn page, which the company sets itself. The second was the LinkedIn company description, read by an AI step. We used both together to place the account into ActiveCampaign’s internal vertical categories rather than LinkedIn’s.

The outcome

Accounts match with the categories the sales team already uses in the CRM. Accounts in the vertical categories ActiveCampaign historically closes at a higher rate get a higher score. Reading the description alongside the tag also catches companies whose LinkedIn category was set by whoever built the page and no longer reflects what the business does.

‍

Scoring the tech stack twice, for two different reasons

What we did

Tech stack detection produced two independent scores. The first covers competitor presence: if the account uses a competitor ActiveCampaign has a strong displacement record against, the score goes up, and if it runs one that is historically hard to replace, the score stays low. The second covers integration presence, where the account runs a tool ActiveCampaign integrates with.

The outcome

Competitor presence carries materially more weight than integration presence, since displacement history is the stronger predictor. An SDR looking at a high-scoring account can see that it is high-scoring because the incumbent is one their team knows how to pitch against.

‍

Building data points that needed custom scraping and retrieval

Several inputs had no clean integration available, so they were assembled:

1. Review volume was pulled from Yelp and Google through native Clay integrations, and from Trustpilot and G2 through custom Google searches run against Claygent
2. All four review counts were aggregated into one figure, with the score scaling on the total rather than on any single site
3. Social following required scraping each account’s website footer to find their Facebook, X, YouTube, and Instagram URLs
4. Claygent then visited each social URL to read the follower count
5. Follower counts were aggregated the same way as reviews, then scored
6. Three binary checks on the website completed the score: public pricing on the site, a free trial, and a newsletter signup

‍

Showing the score where the SDRs already work

What we did

We built a custom HTTP request that syncs every resolved data point and the final score back into ActiveCampaign’s Apollo instance.

The outcome

SDRs never open Clay and we also wanted to avoid SDRs to spend time between switching multiple tools. The score and its underlying data points sit on the account record inside Apollo, where the team already spends its day, and they filter and sort on it there. The sales team was given a working threshold: accounts above 35 to 40 carry enough potential to prioritize.

‍

Why ActiveCampaign Chose Growth Today

A scoring system their own team could use, manage and back up.

ActiveCampaign did not need a list. They needed the research layer underneath their existing SDR motion rebuilt as a system to avoid SDR spend time on manual reseach, using data sources they already paid for,  building in the tool their reps already use in their day to day work.

The build took days rather than the year the manual path required, and it reruns against the full TAM in hours. Growth Today also set the expectation on accuracy directly: scraped and inferred data across 50,000 accounts resolves at roughly 80 to 90% coverage, and some sites cannot be read at all.

• Data sources they already owned. They had existing subscriptions for several data points, so the traffic input added no cost
• Scoring logic they defined. Every input maps to something their sales team believes predicts a win, which was influenced by historical closed won and closed lost data
• Delivered into Apollo. No new tool for the SDR team to learn or check

‍

Results & Impact

RevOps Impact

• 50,000 accounts has an accurate, actionable score, covering the full TAM rather than a worked subset
• 15 data points per account resolve automatically, including several that required custom retrieval
• The research can be reapplied to any account or list, so the scored TAM refreshes instead of decaying
• 80 to 90% field coverage, set as an explicit expectation at the start of the project. Some data points won’t be found by AI.

Sales Impact

• SDRs open Apollo and see the data that drives their actions, with no context switching into a data tool
• A working threshold of 35-40 gives the team a shared definition of a priority account
• Competitor context at the account level tells a rep which incumbent they are pitching against before they write anything
• Account research is automatically done, reps don’t have to spend any time on figuring out what to say and who to chase

Leadership Impact

• The full TAM is ranked, so territory and effort allocation run on data rather than on intuition
• Research capacity moved to selling. The manual path for this work required four SDRs for roughly a year

By scoring the research layer instead of the list, ActiveCampaign turned a 50,000-account TAM into a ranked market its SDRs can work down.

‍